A Framework for Distributed Dynamic Load Balancing in Heterogeneous Cluster

Abstract

Distributed Dynamic load balancing (DDLB) is an important system function destined to distribute workload among available processors to improve throughput and/or execution times of parallel computer in Cluster Computing. Instead of balancing the load in cluster by process migration, or by moving an entire process to a less loaded computer, we make an attempt to balance load by splitting processes into separate jobs and then balance them to nodes. In order to get target, we use mobile agent (MA) to distribute load among nodes in a cluster. In this study, a multi-agent framework for load balancing in heterogeneous cluster is given. Total load on node is calculated using queue length which is measured as the total number of processes in queue. We introduce types of agents along with policies needed to meet the requirements of the proposed load-balancing. Different metrics are used to compare load balancing mechanism with the existing message passing technology. The experiment is carried out on cluster of PC's divided into multiple LAN's using PMADE (Platform for Mobile agent distribution and execution). Preliminary experimental results demonstrated that the proposed framework is effective than the existing ones.